Learning without Human Scores for Blind Image Quality Assessment

By Wufeng Xue, Lei Zhang, Xuanqin Mou. 2013 IEEE Conference on Computer Vision and Pattern Recognition, 2013

402 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/6cfcd4e8e58156703abc863e248ac08b5043b62a

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

MUSIQ: Multi-scale Image Quality Transformer (2021, background): “In the past years, researchers proposed many IQA methods: early natural scene statistics based [13, 25, 27, 45], codebook-based [38, 40] and CNN-based [11, 33, 34, 41, 47].”
Group MAD Competition? A New Methodology to Compare Objective Image Quality Models (2016, background): “The four distortion types are common in existing IQA databases [27, 23] and many IQA models are claimed to be able to properly handle these distortions [19, 20, 25, 17, 40, 18, 38, 37, 26, 42, 36, 7, 35].”
From Patches to Pictures (PaQ-2-PiQ): Mapping the Perceptual Space of Picture Quality (2019, background): “No-reference (NR) or blind picture quality prediction is largely unsolved, though popular models exist [14, 15, 16, 17, 18, 19, 20].”
The Analysis of Image Contrast: From Quality Assessment to Automatic Enhancement (2016, methodology): “images from statistical regularities found in natural images and 2) quality-aware clustering (QAC) [27], using a codebook learnt from a group of quality-aware centroids to”
dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs (2017, methodology): “We compare dipIQ with two well-known FR-IQA models: PSNR (whose largest value is clipped at 60 dB in order to perform a reasonable parameter estimation) and SSIM [94] (whose implementation used in the paper involves a down-sampling process [95]) and previous OU-BIQA models, including QAC [60], NIQE ”
Using Free Energy Principle For Blind Image Quality Assessment (2015, methodology): “We first evaluate the performance of the proposed NFERM with a total number of fifteen IQA approaches on the LIVE database [39]: 1) classical FR PSNR, SSIM [4], MS-SSIM [5]; 2) popular blind distortion-specific WNJE [21], MBBM [22], Sheikh [23], SINE [24], JNB [26], CPBD [27], NFEQM [17]; 3) state-o”
Blind Quality Assessment Based on Pseudo-Reference Image (2018, background): “QAC [41] performs qualityaware clustering and learns a set of quality-aware centroids to act a codebook to infer the quality.”
No-Reference Image Quality Assessment via Transformers, Relative Ranking, and Self-Consistency (2021, background): “The goal of the no-reference image quality assessment (NR-IQA) or blind image quality assessment (BIQA) meth-ods is to provide a solution when the reference image is not available [5, 6, 7, 8].”
MetaIQA: Deep Meta-Learning for No-Reference Image Quality Assessment (2020, methodology): “NR-IQA can be classified into distortion-specific methods [25, 24, 29, 47] and general-purpose methods [39, 32, 56, 51, 50, 12, 54, 49, 10].”
Perceptual Quality Assessment of Smartphone Photography (2020, methodology): “Lastly and perhaps more interestingly, we observe improved performance of quality prediction in MT-S when jointly trained with scene classi-ﬁcation.”
NIMA: Neural Image Assessment (2017, background): “Recently, more complex models such as deep convolutional neural networks (CNNs) have been used to address this problem [4]–[11].”
Blind image quality evaluation using perception based features (2015, methodology): “in [12] proposed a Quality-Aware Clustering (QAC) algorithm that learns a set of quality-aware centroids from pristine and distorted imagery and uses them as codebooks to infer the quality of a test image.”
Mean opinion score (MOS) revisited: methods and applications, limitations and alternatives (2016, methodology): “Although recent novel methods for objective quality assessment [42, 43] have been able to reduce or eliminate the need for using subjective MOS in training (but not validation of course), validation and testing of objective models is still an essential part of the work.”
Convolutional Neural Networks for No-Reference Image Quality Assessment (2014, background): “Previous approaches typically accumulate features over the entire image to obtain statistics for estimating overall quality, and have rarely shown the ability to estimate local quality, except for a simple example in [18].”